JP6300026B2 - Measuring device and measuring method - Google Patents

Description

  The present technology relates to a measurement device and a measurement method, and more particularly, to a measurement device and a measurement method suitable for use when measuring a pulse wave or a pulse.

  In recent years, an optical measurement device that measures a pulse wave or a pulse has become widespread (see, for example, Patent Document 1).

Special table 2008-538186

  However, in an optical measurement device, it is very difficult to remove a body motion component generated by the body motion of a subject to be measured from a measurement signal. In particular, when measurement is performed at a site where the movement of the arm surface or the like is large, the body motion component is often larger than the pulse wave component generated by the pulse wave, making it more difficult to remove the body motion component. As a result, the measurement accuracy of the pulse wave and the pulse is lowered.

  Therefore, the present technology is intended to improve the measurement accuracy of pulse waves and pulses.

  The measuring device of one side of this art is from the 1st zone signal containing the component of the 1st frequency band of the 1st measurement signal acquired by irradiating the part containing light with the light of the 1st wavelength. A body motion signal extraction unit that extracts a body motion signal including a component caused by body motion, and a difference between the second band signal including the component of the second frequency band of the first measurement signal and the body motion signal And a calculation unit that generates a pulse wave signal as a signal.

  The body motion signal extraction unit may predict and extract the body motion signal using an autoregressive model.

  The body motion signal extraction unit can generate the autoregressive model within the range from the 5th order to the 12th order using the Yule-Walker method.

  The body motion signal extraction unit can set the order of the autoregressive model based on the level of the first band signal.

  A body motion detector that detects the body motion, and a measurement that measures a pulse wave frequency based on a signal selected from the pulse wave signal and the second band signal based on the detection result of the body motion Can be further provided.

  The body motion signal extraction unit extracts the body motion signal from a signal obtained by attenuating a frequency component of a band including a measurement value of a previous pulse wave frequency among frequency components of the first band signal. Can do.

  The measurement unit includes a frequency detection unit that detects a first peak frequency that is a peak frequency of the pulse wave signal and a second peak frequency that is a peak frequency of the second band signal, and the body motion And a selection unit that selects the pulse wave frequency from the first peak frequency and the second peak frequency based on at least one of the detection result of the above and the measurement value of the previous pulse wave frequency. it can.

  The frequency detection unit can limit a frequency band in which the first peak frequency and the second peak frequency are detected based on a previous measurement value of the pulse wave frequency.

  The frequency detection unit is configured to detect the first peak frequency based on a result of Fourier transform of the pulse wave signal obtained by padding a sample having a value of 0, and to pad the sample having a value of 0. The second peak frequency can be detected based on the result of Fourier transform of the band signal.

  The measurement unit includes a selection unit that selects the pulse wave signal or the second band signal based on the detection result of the body movement, and the peak frequency of the signal selected by the selection unit is the pulse wave frequency. And a frequency detection unit for detecting as follows.

  The frequency detection unit can limit a frequency band in which the peak frequency is detected based on a previous measurement value of the pulse wave frequency.

  The frequency detection unit may detect the peak frequency based on a result of Fourier transform of a signal obtained by padding a signal selected by the selection unit with a sample having a value of 0.

  The body motion detection unit can detect the body motion based on a frequency distribution of the body motion signal.

  The body motion detection unit includes a third band signal including a component of the first frequency band of the second measurement signal acquired by irradiating the portion including the pulse with light of the second wavelength; The body motion can be detected based on the distribution of the combined vector with the first band signal.

  Based on the change in the first measurement signal and the change in the second measurement signal obtained by irradiating the portion including the pulse with the light of the second wavelength in the body motion detection unit, The body movement can be detected.

  The measurement unit can calculate the pulse rate based on the pulse wave frequency.

  A first filter for extracting the first band signal from the first measurement signal, and a second filter for extracting the second band signal from the first measurement signal; In this frequency band, the range of the pulse wave frequency to be measured can be included, and the maximum value of the second frequency band can be made larger than the maximum value of the first frequency band.

  A filter for extracting the first band signal from the first measurement signal is further provided, the first frequency band is the same as the second frequency band, and includes a range of pulse wave frequencies to be measured. Thus, the first band signal and the second band signal can be the same signal.

  A measurement method according to one aspect of the present technology includes a first band signal including a component of a first frequency band of a first measurement signal acquired by irradiating a portion including a pulse with light having a predetermined wavelength. A body motion signal extraction step for extracting a body motion signal including a component caused by body motion, and a difference signal between the second band signal including a component of the second frequency band of the first measurement signal and the body motion signal And generating a pulse wave signal.

  In one aspect of the present technology, from the first band signal including the component of the first frequency band of the first measurement signal acquired by irradiating the portion including the pulse with light having a predetermined wavelength, the body movement is performed. Is extracted, and a pulse wave signal, which is a difference signal between the second band signal including the component of the second frequency band of the first measurement signal and the body movement signal, is generated. The

  According to one aspect of the present technology, a pulse wave component can be extracted from a measurement signal with high accuracy. In addition, according to one aspect of the present technology, the measurement accuracy of a pulse wave and a pulse can be improved.

It is a figure for demonstrating the relationship between a pulse wave and a body motion. It is an outline view showing one embodiment of a measuring device to which this art is applied. It is a block diagram which shows the structural example of the main-body part of a measuring device. It is a block diagram which shows the structural example of light reception IC. It is a timing chart for demonstrating the operation example of light reception IC. It is a block diagram which shows 1st Embodiment of the arithmetic processing part of a measuring device. It is a figure for demonstrating the wavelength of measurement light. It is a flowchart for demonstrating 1st Embodiment of a pulse measurement process. It is a figure which shows the example of the detection range of a peak frequency. It is a graph which shows the example of a measurement waveform. It is a figure for demonstrating the modification of the ratio of downsampling. It is a block diagram which shows 2nd Embodiment of the arithmetic processing part of a measuring device. It is a flowchart for demonstrating 2nd Embodiment of a pulse measurement process. It is a graph which shows the 1st example of the measurement signal with respect to the measurement light of each wavelength. It is a graph which shows the 2nd example of the measurement signal with respect to the measurement light of each wavelength. It is a figure which shows the example of distribution of a synthetic vector. It is a block diagram which shows 3rd Embodiment of the arithmetic processing part of a measuring device. It is a flowchart for demonstrating 3rd Embodiment of a pulse measurement process. It is a graph which shows the example of a measurement waveform. It is a graph which shows the example of the envelope of the measurement signal before band limitation and after limitation It is a figure which shows the modification of the detection range of a peak frequency. It is a block diagram which shows the structural example of a computer.

Hereinafter, modes for carrying out the present technology (hereinafter referred to as embodiments) will be described. The description will be given in the following order.
1. 1. Relationship between pulse wave and body movement First Embodiment 3. FIG. Second embodiment 4. Third embodiment 5. Modified example

<1. Relationship between pulse wave and body movement>
Each graph of FIG. 1 shows an example of a result of frequency analysis of a signal obtained by measuring a pulse wave of a subject's arm with an optical measurement device by FFT (Fast Fourier Transform). The horizontal axis of each graph indicates the frequency, and the vertical axis indicates the spectrum intensity. In addition, the change in the frequency distribution of the measurement signal when the subject steps from a stationary state and then starts running is shown in order from the top graph.

  The top graph in FIG. 1 shows the frequency distribution of the measurement signal when the subject is stationary. The second to fourth graphs from the top show the frequency distribution of the measurement signal when the subject is stepping on. The fifth to eighth graphs from the top show the analysis results when the subject is running. Note that the pulse of the subject when measuring the measurement signal of each graph is 68 bpm (beats per minute), 89 bpm, 84 bpm, 86 bpm, 89 bpm, 94 bpm, 94 bpm, and 102 bpm in order from the top.

  Moreover, the peak of the frequency indicated by the arrow in each graph indicates the frequency of the subject's pulse wave (hereinafter referred to as pulse wave frequency), and the peak of the frequency indicated by the black circle indicates the frequency of the subject's body movement (hereinafter, (Referred to as body motion frequency).

  As shown in this example, the fundamental frequency of the body motion component is generally slightly lower than the pulse wave frequency. The body motion component includes a harmonic component with high periodicity. Furthermore, the energy of the body motion component is very large compared to the energy of the pulse wave component. Further, the frequency spectrum of the pulse wave component and the frequency spectrum of the body motion component approach as the momentum increases and the pulse becomes faster.

  Therefore, it is very difficult to accurately separate the pulse wave component and the body motion component with a comb filter or a BPF (band pass filter). Further, since the energy of the body motion component is very large compared to the energy of the pulse wave component, it is very difficult to learn the pulse wave at rest and extract it from the measurement signal during body motion. Further, in order to extract the pulse wave component from the measurement signal with higher accuracy, a huge amount of calculation is required. For this reason, resources such as CPU (Central Processing Unit) and memory and power consumption increase, processing time becomes long, and real-time processing becomes difficult.

  In each embodiment of the present technology described below, it is possible to remove the influence of the body motion component having such a property and to measure the pulse wave and the pulse with high accuracy.

<2. First Embodiment>
{Configuration example of measuring device 1}
2 and 3 show a configuration example of the measurement apparatus 1 that is the first embodiment of the measurement apparatus to which the present technology is applied. FIG. 2 shows an external configuration example of the measuring apparatus 1. FIG. 3 shows a configuration example of the main body 11 of the measuring apparatus 1.

  The measuring device 1 is a wristband type measuring device that measures a subject's pulse by an optical method. As shown in FIG. 2, the measuring device 1 includes a main body 11 and a band 12, and the band 12 is attached to the arm (wrist) 2 of the subject like a wristwatch. And the main-body part 11 irradiates the part containing the pulse of the test subject's arm 2 with the measurement light of a predetermined wavelength, and measures a test subject's pulse based on the intensity | strength of the returned light.

  The main body 11 is configured to include a substrate 21, an LED 22, a light receiving IC 23, a light shield 24, an operation unit 25, an arithmetic processing unit 26, a display unit 27, and a wireless device 28. The LED 22, the light receiving IC 23, and the light shield 24 are provided on the substrate 21.

  The LED 22 irradiates the portion including the pulse of the arm 2 of the subject under the control of the light receiving IC 23.

  The light receiving IC 23 receives light that has returned after the measurement light is applied to the arm 2. The light receiving IC 23 generates a digital measurement signal indicating the intensity of the returned light, and supplies the generated measurement signal to the arithmetic processing unit 26.

  The light shield 24 is provided between the LED 22 and the light receiving IC 23 on the substrate 21. The light shield 24 prevents the measurement light from the LED 22 from directly entering the light receiving IC 23.

  The operation unit 25 is composed of various operation members such as buttons and switches, and is provided on the surface of the main body unit 11. The operation unit 25 is used to operate the measuring device 1 and supplies a signal indicating the operation content to the arithmetic processing unit 26.

  The arithmetic processing unit 26 performs arithmetic processing for measuring the pulse of the subject based on the measurement signal supplied from the light receiving IC 23. The arithmetic processing unit 26 supplies the pulse measurement result to the display unit 27 and the wireless device 28.

  The display unit 27 is configured by a display device such as an LCD (Liquid Crystal Display), and is provided on the surface of the main body unit 11. The display unit 27 displays the measurement result of the subject's pulse.

  The wireless device 28 transmits the measurement result of the subject's pulse to an external device by wireless communication of a predetermined method. For example, as illustrated in FIG. 3, the wireless device 28 transmits the measurement result of the subject's pulse to the smartphone 3 and causes the screen 3 </ b> A of the smartphone 3 to display the measurement result. Note that any method can be adopted as the communication method of the wireless device 28.

{Configuration example of light receiving IC 23}
FIG. 4 shows a configuration example of the light receiving IC 23. In this example, an example in which the LED 22 of the measuring device 1 includes three LEDs, LEDs 22a to 22c, is shown. The LEDs 22a to 22c emit measurement light having different wavelengths.

  The light receiving IC 23 is configured to include an LED driver 51, a selector 52, a light receiving element 53, an AD converter 54, and a communication unit 55.

  The LED driver 51 supplies drive signals to the LEDs 22a to 22c via the selector 52, and controls lighting, extinction, light emission amount, and the like of the LEDs 22a to 22c.

  The selector 52 selects the supply destination of the drive signal supplied from the LED driver 51 from the LEDs 22a to 22c, and supplies the drive signal to the selected LED.

  The light receiving element 53 receives light that returns after the measurement light from the LEDs 22a to 22c is applied to the arm 2. The light receiving element 53 supplies a measurement signal, which is an analog electric signal indicating the intensity of the received light, to the AD converter 54.

  The AD converter 54 samples the measurement signal at a predetermined sampling frequency, and converts the analog measurement signal into a digital measurement signal. The AD converter 54 supplies a digital measurement signal to the communication unit 55. Note that the sampling frequency of the AD converter 54 is, for example, 200 to 220 Hz. Hereinafter, a case where the sampling frequency of the AD converter 54 is 200 Hz will be described as an example.

  The communication unit 55 communicates with the arithmetic processing unit 26 by wired communication of a predetermined method, and transmits and receives measurement signals and various control signals. Note that an arbitrary method such as I2C can be adopted as a communication method of the communication unit 55.

  As shown in FIG. 4, the light receiving IC 23 can also be used when measuring a pulse in a part other than the subject's arm 2 (for example, a finger 4, an earlobe (not shown), etc.).

{Operation example of light receiving IC 23}
Here, an operation example of the light receiving IC 23 will be described with reference to the timing chart of FIG. The first row in FIG. 5 shows the timing of the control signal supplied from the arithmetic processing unit 26 to the communication unit 55 of the light receiving IC 23. The second row shows the operation period of the AD converter 54. The third row shows the light emission period of the LEDs 22a to 22c. In this figure, the LED 22a is represented as LEDa, the LED 22b is represented as LEDb, and the LED 22c is represented as LEDc.

  The light receiving IC 23 starts its operation when a write signal is supplied from the arithmetic processing unit 26 to the communication unit 55.

  Specifically, first, the LED driver 51 supplies a drive signal to the LED 22a via the selector 52, and causes the LED 22a to emit pulsed measurement light. The light receiving element 53 receives the light returned after the measurement light from the LED 22a is irradiated on the arm 2. The light receiving element 53 supplies a measurement signal (hereinafter referred to as a measurement signal a), which is an analog electric signal indicating the intensity of received light, to the AD converter 54. The AD converter 54 samples the analog measurement signal a and performs AD conversion to the digital measurement signal a. The AD converter 54 supplies the digital measurement signal a to the communication unit 55.

  The LED driver 51 supplies a drive signal to the LED 22b via the selector 52 after the AD conversion of the measurement signal a is completed, and causes the LED 22b to emit pulsed measurement light. The light receiving element 53 receives the light returned after the measurement light from the LED 22b is applied to the arm 2. The light receiving element 53 supplies a measurement signal (hereinafter referred to as a measurement signal b) which is an analog electric signal indicating the intensity of the received light to the AD converter 54. The AD converter 54 samples the analog measurement signal b and performs AD conversion to the digital measurement signal b. The AD converter 54 supplies the digital measurement signal b to the communication unit 55.

  The LED driver 51 supplies a drive signal to the LED 22c via the selector 52 after the AD conversion of the measurement signal b is completed, and causes the LED 22c to emit pulsed measurement light. The light receiving element 53 receives the light returned after the measurement light from the LED 22c is irradiated on the arm 2. The light receiving element 53 supplies a measurement signal (hereinafter referred to as a measurement signal c) that is an analog electric signal indicating the intensity of the received light to the AD converter 54. The AD converter 54 samples the analog measurement signal c and performs AD conversion to the digital measurement signal c. The AD converter 54 supplies the digital measurement signal c to the communication unit 55.

  The communication unit 55 supplies the measurement signals a to c to the arithmetic processing unit 26 when the read signal is supplied from the arithmetic processing unit 26.

  Thereafter, after an idle period of a predetermined time elapses, a series of processes from the light emission of the LED 22a to the supply of the measurement signals a to c to the arithmetic processing unit 26 are performed.

  The above process is repeatedly executed, for example, until a predetermined period or number of times or a stop command is input from the arithmetic processing unit 26.

  4 and 5 show examples in which there are three types of wavelengths of measurement light, it is also possible to set the wavelength of measurement light to one type, two types, or four or more types. Hereinafter, in the first embodiment, a case where the wavelength of the measurement light is one type will be described as an example.

{Configuration example of arithmetic processing unit 26a}
FIG. 6 shows a configuration example of the arithmetic processing unit 26a which is the first embodiment of the arithmetic processing unit 26 of the measuring device 1. The arithmetic processing unit 26a includes a decimation filter 101, a BPF (bandpass filter) 102, an autocovariance function estimation unit 103, a linear prediction filter 104, a BPF (bandpass filter) 105, an arithmetic unit 106, and a DFT (discrete Fourier transform) unit. 107a and 107b, peak detection units 108a and 108b, DFT (Discrete Fourier Transform) unit 109, determination unit 110, selection unit 111, calculation unit 112, and storage unit 113.

  The body motion signal extraction unit 131 is configured by the self-covariance function estimation unit 103 and the linear prediction filter 104. The BPF 102, the BPF 105, the calculation unit 106, and the body motion signal extraction unit 331 constitute a pulse wave signal extraction unit 132. The DFT unit 109 and the determination unit 110 constitute a body motion detection unit 133. A frequency detection unit 134 is configured by the DFT units 107a and 107b and the peak detection units 108a and 108b. The selection unit 111, the calculation unit 112, the storage unit 113, and the frequency detection unit 134 constitute a measurement unit 135.

  The decimation filter 101 down-samples the measurement signal. The decimation filter 101 supplies the measurement signal after downsampling to the BPF 102 and the BPF 105.

  The BPF 102 is composed of, for example, a zero phase filter composed of a two-stage third-order IIR (Infinite Impulse Response) filter. The BPF 102 passes components in a predetermined frequency band (hereinafter referred to as frequency band N) of the measurement signal and blocks components other than the frequency band N. For example, the BPF 102 applies the third-order IIR filter of the first stage to the measurement signal, and then reverses the sample order to the second-order third-order IIR that is the same as the first stage. The phase distortion is canceled by a method such as applying a filter. The BPF 102 supplies a signal including the extracted component of the frequency band N (hereinafter referred to as band signal N) to the self-covariance function estimation unit 103 and the linear prediction filter 104.

  Since the BPF 102 is a zero phase filter, time axis information such as a body motion component is held in the extracted band signal N.

  The autocovariance function estimation unit 103 estimates the autocovariance function of the body motion signal included in the band signal N. The body motion signal is a signal including a body motion component generated by the body motion of the subject. The autocovariance function estimation unit 103 supplies the estimation result of the autocovariance function to the linear prediction filter 104.

  The linear prediction filter 104 includes, for example, a linear prediction filter created by a power spectral density estimation algorithm based on the Yule-Walker method. For example, the linear prediction filter 104 obtains the parameters of the AR model (autocorrelation model) of the body motion signal by the Yule-Walker method using the autocovariance function estimated by the autocovariance function estimation unit 103. Thereby, an AR model of the body motion signal is generated. The linear prediction filter 104 predicts a body motion signal using the generated AR model, and when the body motion component is sufficiently larger than the pulse wave component, only the body motion component that does not include a low-level pulse wave component and a noise component. Is extracted from the band signal N. The linear prediction filter 104 supplies the extracted body motion signal to the calculation unit 106 and the DFT unit 109.

  For example, the BPF 105 includes a zero-phase filter constituted by a two-stage third-order IIR (Infinite Impulse Response) filter, like the BPF 102. The BPF 105 passes components in a predetermined frequency band (hereinafter referred to as frequency band W) of the measurement signal and blocks components other than the frequency band W. The BPF 105 supplies a signal including the extracted component of the frequency band W (hereinafter referred to as a band signal W) to the arithmetic unit 106 and the DFT unit 107b.

  Since the BPF 105 is a zero phase filter, time axis information such as a pulse wave component is held in the extracted band signal W.

  The calculation unit 106 adds the band signal W and the inverse signal of the body motion signal to obtain a difference between the band signal W and the body motion signal. The calculation unit 106 supplies a differential signal (hereinafter referred to as a pulse wave signal) generated by taking the difference between the band signal W and the body motion signal to the DFT unit 107a.

  The DFT unit 107a performs DFT of the pulse wave signal and supplies the analysis result of the frequency of the pulse wave signal to the peak detection unit 108a.

  The peak detector 108a detects the peak frequency of the pulse wave signal based on the frequency analysis result of the pulse wave signal. At this time, the peak detection unit 108a limits the frequency band in which the peak frequency is detected based on the previous detection value of the pulse wave frequency stored in the storage unit 113. The peak detection unit 108 a supplies the detection value of the peak frequency of the pulse wave signal to the selection unit 111.

  The DFT unit 107b performs DFT of the band signal W, and supplies the analysis result of the frequency of the band signal W to the peak detection unit 108b.

  The peak detection unit 108b detects the peak frequency of the band signal W based on the frequency analysis result of the band signal W. At this time, the peak detection unit 108b limits the frequency band in which the peak frequency is detected based on the previous detection value of the pulse wave frequency stored in the storage unit 113. The peak detection unit 108 b supplies the detection value of the peak frequency of the band signal W to the selection unit 111.

  The DFT unit 109 performs DFT of the band signal N and supplies the analysis result of the frequency of the band signal N to the determination unit 110.

  Based on the frequency analysis result of the band signal N, the determination unit 110 determines the presence / absence of a body movement (hereinafter referred to as an obstacle body movement) that is an obstacle to pulse measurement. The determination unit 110 supplies the determination result of the presence or absence of obstacle body movement to the selection unit 111.

  The selection unit 111 selects the pulse wave frequency from the peak frequency of the pulse wave signal and the peak frequency of the band signal W based on the determination result of the presence or absence of obstacle body movement and the previous measurement value of the pulse wave frequency. . The selection unit 111 supplies information indicating the selected pulse wave frequency to the calculation unit 112 and causes the storage unit 113 to store the information.

  The calculation unit 112 calculates the pulse rate based on the pulse wave frequency. The calculation unit 112 outputs the calculated pulse rate to the outside as a measurement result.

  The memory | storage part 113 memorize | stores the measured value of the past pulse wave frequency.

{Measurement light wavelength}
FIG. 7 shows absorption characteristics for light in each wavelength band of Hb (reduced hemoglobin) and HbO ₂ (oxygenated hemoglobin) contained in blood. The horizontal axis indicates the wavelength, and the vertical axis indicates the absorption coefficient. A curve 151 indicates the absorption characteristic of Hb, and a curve 152 indicates the absorption characteristic of HbO ₂ .

For example, light having a wavelength of 470 nm (hereinafter referred to as blue measurement light) or light having a wavelength of 660 nm (hereinafter referred to as red measurement light) having a large difference in absorption coefficient between Hb and HbO ₂ in the visible light band is measured. Used for light. Further, for example, the absorption coefficient of Hb and HbO ₂ in the visible light band is substantially equal wavelength of 530nm light (hereinafter, referred to as green measurement light) and a wavelength 585nm of light (hereinafter, referred to as yellow measurement light) is measured Used for light. Further, for example, light having a wavelength of approximately 805 nm where the absorption coefficients of Hb and HbO ₂ are approximately equal in the infrared light band, or light having a wavelength of 880 nm having a large difference between the absorption coefficients of Hb and HbO ₂ is used as the measurement light.

{First embodiment of pulse measurement processing}
Next, a first embodiment of the pulse measurement process executed by the measurement device 1 will be described with reference to the flowchart of FIG.

  The pulse is measured at a predetermined interval (for example, every 8 to 15 seconds). Hereinafter, a case where pulse measurement is performed at intervals of 8 seconds will be described as an example.

  In step S1, the measurement apparatus 1 starts acquiring measurement signals. That is, as described above with reference to FIG. 5, light emission by the LED 22 and light reception by the light receiving IC 23 are started. In addition, supply of the measurement signal from the light receiving IC 23 to the arithmetic processing unit 26 of the measurement signal indicating the intensity of the received light is started.

  In step S2, the decimation filter 101 down-samples the measurement signal in a band necessary for analyzing the pulse wave signal. For example, the decimation filter 101 performs decimation processing on the measurement signal, and down-samples the number of samples of the measurement signal at a predetermined ratio. The decimation filter 101 supplies the downsampled measurement signal to the BPF 102 and the BPF 105.

  In the present example, since the sampling frequency of the measurement signal is 200 Hz and one measurement period is 8 seconds, the number of samples of the measurement signal in one measurement period is 1600 samples. For example, when the downsampling ratio is 1/16, the number of measurement signal samples in one measurement period decreases from 1600 samples to 100 samples.

  In this case, the sampling frequency of the measurement signal after downsampling is 12.5 Hz (= 200 Hz × 1/16). Therefore, it is possible to detect up to a frequency component of 6.25 Hz using the measurement signal after downsampling, as indicated by the range R1 in FIG.

  On the other hand, the maximum value of the human pulse rate is about 220 bpm, and it is considered sufficient that the maximum value of the pulse rate that can be measured by the measuring apparatus 1 is 240 bpm. When the pulse rate of 240 bpm is converted into a pulse wave frequency, 4.0 Hz is obtained. Therefore, if it is possible to detect a frequency component of 6.25 Hz from the measurement signal after downsampling, it is possible to sufficiently measure the human pulse wave frequency.

  In addition, the subsequent calculation amount can be reduced by down-sampling the measurement signal.

  Hereinafter, a case where the downsampling ratio is set to 1/16 will be described.

  In step S3, the pulse wave signal extraction unit 132 limits the frequency band of the measurement signal. Specifically, the BPF 102 extracts a component in the frequency band N of the measurement signal after downsampling. The BPF 102 supplies the band signal N including the extracted frequency band N component to the self-covariance function estimation unit 103 and the linear prediction filter 104.

  The BPF 105 extracts the frequency band W component of the measurement signal after downsampling. The BPF 105 supplies the band signal W including the extracted band signal W component to the calculation unit 106 and the DFT unit 107b.

  Here, the frequency band N and the frequency band W are set according to the assumed ranges of the pulse wave frequency and the body motion frequency. That is, the frequency band N and the frequency band W are set according to the pulse wave frequency range to be measured and the body motion component frequency range to be extracted.

  For example, the frequency band W is set so as to include at least the range of the pulse wave frequency to be measured. On the other hand, the frequency band N is set so as to include at least the frequency range of the body motion component assumed to be included in the frequency band W, for example.

  For example, when the measurable range of the pulse rate of the measuring device 1 is 30 bpm to 240 bpm, when this range is converted into a pulse wave frequency, it becomes 0.5 Hz to 4.0 Hz. In this case, the minimum value of the frequency band W is set to 0.5 Hz or less, and the maximum value is set to 4.0 Hz or more.

  On the other hand, as described above with reference to FIG. 1, the fundamental frequency of the body motion component tends to be slightly lower than the pulse wave frequency. In this case, assuming that the range of the fundamental frequency of the body motion component to be detected is 0.5 Hz to 2.5 Hz, the minimum value of the frequency band N is set to 0.5 Hz or less, and the maximum value is set to 2.5 Hz or more. Is set.

  Hereinafter, a case where the frequency band W is set to 0.5 Hz to 4.0 Hz and the frequency band N is set to 0.5 Hz to 2.5 Hz will be described. That is, in this case, the frequency band W is set to a band wider than the frequency band N. More specifically, the maximum value of the frequency band W is set to a value larger than the maximum value of the frequency band N, and the minimum value of the frequency band W is set to the same value as the minimum value of the frequency band N.

  In step S4, the body motion signal extraction unit 131 extracts a body motion signal. For example, the auto-covariance function estimation unit 103 includes a body included in the band signal N based on a predetermined number (for example, eight) of samples at the head of 100 samples within one measurement period of the band signal N. Estimate the autocovariance function of the dynamic signal. The autocovariance function estimation unit 103 supplies the estimation result of the autocovariance function to the linear prediction filter 104.

  The linear prediction filter 104 uses the estimated auto-covariance function to calculate the body based on a predetermined number (for example, 8) samples at the beginning of 100 samples within one measurement period of the band signal N. The parameters of the AR model of the dynamic signal are obtained by the Yule-Walker method. Thereby, an AR model of the body motion signal is generated. The linear prediction filter 104 predicts a body motion signal using the generated AR model. The linear prediction filter 104 removes the pulse wave component and the noise component from the remaining predetermined number (for example, 92) of samples in one measurement period of the band signal N based on the prediction result of the body motion signal, Extract body motion signals.

  Here, the order of the AR model of the body motion signal will be described. When only the noise component is to be accurately removed from the band signal N, it is desirable to set the order of the AR model to about 20th to 30th, for example. However, in this case, not only the body motion component but also the pulse wave component are extracted from the band signal N.

  On the other hand, when the order of the AR model is kept low, the accuracy of removing the noise component of the linear prediction filter 104 decreases. On the other hand, when the band signal N contains many body motion components, not only the noise component but also a pulse wave component that is very weak compared to the body motion component can be removed. As a result, there is an effect that only the body motion component can be extracted from the band signal N with a small amount of calculation.

  Therefore, the order of the AR model is set to 8th order, for example. The order of the eighth order is a value obtained by experiments as an order capable of extracting the body motion component from the band signal N with high accuracy. However, the order of the AR model is not necessarily limited to the 8th order. For example, the body motion component can be extracted from the band signal N with high accuracy even if the order of the AR model is set within the range of the fifth order to the twelfth order.

  The linear prediction filter 104 supplies the extracted body motion signal to the calculation unit 106 and the DFT unit 109.

  When the body motion component is not included in the band signal N, or when the body motion component included in the band signal N is small, the pulse wave component is also extracted from the band signal N by the linear prediction filter 104 and is converted into the body motion signal. A pulse wave component is included.

  In step S5, the calculation unit 106 extracts a pulse wave signal. Specifically, the operation unit 106 adds the band signal W and the inverse signal of the body motion signal to obtain a difference between the band signal W and the body motion signal, and generates a pulse wave signal that is a difference signal. The calculation unit 106 supplies the pulse wave signal to the DFT unit 107a.

  The uppermost graph in FIG. 10 shows an example of the measurement result (solid line) of the frequency distribution of the band signal W and the prediction result (dotted line) of the frequency distribution of the body motion signal using the AR model. The horizontal axis of the graph indicates the frequency, and the vertical axis indicates the spectral intensity.

  The second graph from the top in FIG. 10 shows an example of the measurement result (solid line) of the waveform of the band signal W and the prediction result (dotted line) of the waveform of the body motion signal using the AR model. The horizontal axis of the graph indicates time, and the vertical axis indicates amplitude.

  The third graph from the top in FIG. 10 shows an example of a pulse wave signal that is a difference signal between the band signal W (solid line) and the body motion signal (dotted line) in the second graph from the top. The horizontal axis of the graph indicates time, and the vertical axis indicates amplitude.

  The fourth graph from the top in FIG. 10 shows an example of the frequency distribution of the band signal W (solid line) in the second graph from the top. The horizontal axis of the graph indicates the frequency, and the vertical axis indicates the spectral intensity.

  The bottom graph of FIG. 10 shows an example of the frequency distribution of the pulse wave signal in the third graph from the top. The horizontal axis of the graph indicates the frequency, and the vertical axis indicates the spectral intensity.

  As shown in this example, by taking the difference between the band signal W and the body motion signal, the body motion component is removed from the band signal W, and the pulse wave signal including the pulse wave component is extracted. That is, the body motion component can be canceled efficiently.

  As described above, when the body motion component is not included in the band signal N, or when the body motion component included in the band signal N is small, the body motion signal includes a pulse wave component. Therefore, the pulse wave component, which is the difference signal between the band signal W and the body motion signal, hardly contains the pulse wave component.

  In step S6, the frequency detector 134 detects the peak frequency of the pulse wave signal. Specifically, the DFT unit 107a pads a predetermined number of samples of value 0 between each sample of 100 samples of pulse wave signals within one measurement period, and up-samples them to 1024 samples. Then, the DFT unit 107a performs DFT of the pulse wave signal after padding, and supplies the analysis result of the frequency of the pulse wave signal to the peak detection unit 108a.

  It should be noted that the resolution of the frequency analysis of the pulse wave signal can be increased by padding the sample of the value 0, setting the number of samples of the pulse wave signal to 1024 or 2048 samples, and increasing the sampling frequency. Also, by padding samples with a value of 0, DFT can be performed using a sparse matrix, and the computation processing can be speeded up.

  The peak detector 108a detects the peak frequency of the pulse wave signal based on the frequency analysis result of the pulse wave signal. At this time, the peak detection unit 108 a reads the previous measurement value of the pulse wave frequency from the storage unit 113. And the peak detection part 108a restrict | limits the frequency band which detects a peak frequency based on the measured value of the last pulse wave frequency.

  For example, when the measurement value of the previous pulse wave frequency is 2 Hz, the peak detection unit 108a detects a predetermined bandwidth centering on 2 Hz as a frequency band for detecting the peak frequency as shown in FIG. Limited to range R2. The detection range R2 is set to a range narrower than the bandwidth R3 of the frequency band W.

  The peak detector 108a detects the peak frequency of the pulse wave signal within the set detection range. The peak detection unit 108a supplies the detection value of the peak frequency to the selection unit 111.

  In this way, by detecting the peak frequency by limiting the detection range based on the previous measurement value of the pulse wave frequency, the calculation amount is reduced, and a peak frequency different from the pulse wave frequency may be detected. Is reduced.

  In step S7, the frequency detector 134 detects the peak frequency of the measurement signal. Specifically, the DFT unit 107b performs DFT of the band signal W after padding the band signal W with a sample of value 0 by the same processing as the DFT unit 107a in step S6. Further, the peak detection unit 108b detects the peak frequency of the band signal W after limiting the detection range based on the previous measurement value of the pulse wave frequency by the same processing as the peak detection unit 108a in step S6. Thereby, the peak frequency in the detection range of the measurement signal is detected. The peak detection unit 108 b supplies the peak frequency detection value to the selection unit 111.

  In step S8, the body movement detection unit 133 detects body movement. Specifically, the DFT unit 109 performs DFT of the body motion signal and supplies the analysis result of the frequency of the body motion signal to the determination unit 110.

  The determination unit 110 determines the presence / absence of an obstruction body motion based on the frequency distribution of the body motion signal. For example, when the frequency range of the body motion signal is greater than or equal to a predetermined width, the determination unit 110 determines that there is a disabled body motion. On the other hand, when the frequency range of the body motion signal is less than a predetermined width, the determination unit 110 determines that there is no obstacle body motion.

  Here, the frequency range of the body motion signal is obtained, for example, by the difference between the minimum frequency and the maximum frequency among the frequencies having the spectrum intensity of the body motion signal equal to or higher than a predetermined threshold.

  Alternatively, for example, the determination unit 110 determines the presence / absence of obstacle body movement based on the frequency distribution waveform of the body movement signal. For example, the determination unit 110 uses a discriminator obtained by prior machine learning, and the waveform of the frequency distribution of the body motion signal includes a disability body motion component (a body motion component that impedes pulse measurement). Or not. If the determination unit 110 determines that the obstruction body motion component is included in the body motion signal as a result of the identification, the determination unit 110 determines that there is an obstruction body motion. On the other hand, when the determination unit 110 determines that the obstacle body motion component is not included in the body motion signal as a result of the identification, it determines that there is no obstacle body motion.

  Then, the determination unit 110 supplies the determination unit 111 with the determination result of the presence or absence of obstacle body movement.

  Note that the criterion for determining whether or not the subject's body movement is an obstacle body movement is set by, for example, prior learning or experiment. By appropriately setting this criterion, weak body movements that do not interfere with pulse measurement are ignored.

  In step S <b> 9, the selection unit 111 selects a pulse wave frequency from the detected peak frequency based on the detection result of the body motion and the previous measurement value of the pulse wave frequency. For example, the selection unit 111 reads the previous measurement value of the pulse wave frequency from the storage unit 113. Based on the measurement value of the previous pulse wave frequency, the selection unit 111 sets the maximum range in which the pulse wave frequency is assumed to be changeable between the previous measurement and the current measurement as the selection reference range. Set.

  Then, when only one of the peak frequency of the pulse wave signal and the peak frequency of the band signal W is included in the selection reference range, the selection unit 111 selects the peak frequency included in the selection reference range as the pulse wave. Select frequency.

  On the other hand, when both peak frequencies are included in the selection reference range, or when both peak frequencies are not included in the selection reference range, the selection unit 111 is based on the determination result of the presence or absence of obstacle movement. To select the pulse wave frequency.

  Specifically, as described above, when there is an obstruction body motion and a lot of body motion components are included in the measurement signal, almost no pulse wave component remains in the body motion signal. Therefore, the pulse wave component, which is the difference signal between the band signal W and the body motion signal, includes a pulse wave component, but hardly includes the body motion component. As a result, it is assumed that the peak frequency of the pulse wave signal is substantially equal to the pulse wave frequency of the subject. Therefore, the selection unit 111 selects the peak frequency of the pulse wave signal as the pulse wave frequency when it is determined that there is an obstacle movement. That is, the peak frequency of the pulse wave signal becomes the measured value of the current pulse wave frequency.

  On the other hand, as described above, when there is no obstruction body motion and the body motion component is not included in the measurement signal, the pulse wave component remains in the body motion signal without being removed. Accordingly, the pulse wave component, which is the difference signal between the band signal W and the body motion signal, contains almost no pulse wave component. On the other hand, the body signal is hardly included in the band signal W. Therefore, it is assumed that the peak frequency of the band signal W is substantially equal to the pulse wave frequency of the subject. Therefore, the selection unit 111 selects the peak frequency of the band signal W as the pulse wave frequency when it is determined that there is no obstacle movement. That is, the peak frequency of the band signal W is the measured value of the current pulse wave frequency.

  The selection unit 111 supplies information indicating the selected pulse wave frequency to the calculation unit 112. Further, the selection unit 111 causes the storage unit 113 to store information indicating the selected pulse wave frequency. Thereby, the measurement value of the current pulse frequency is stored in the storage unit 113.

  In step S10, the calculation unit 112 calculates the pulse rate. For example, the calculation unit 112 obtains the pulse rate by multiplying the pulse wave frequency selected by the selection unit 111 by 60 times.

  In step S11, the calculation unit 112 outputs a measurement result. That is, the calculation part 112 outputs the pulse rate calculated | required by the process of step S10 as a measurement result outside.

  Thereafter, the process returns to step S2, and the processes after step S2 are repeatedly executed.

  In this way, it is possible to accurately measure the pulse wave and pulse of the subject by removing the influence of body movement with a small amount of calculation. For example, even if the subject performs intense exercise such as running, the pulse wave and pulse of the subject can be accurately measured. Further, for example, even when the subject performs measurement while wearing the measuring device 1 for a long time, the influence of the subject's body movement can be removed and the pulse wave and the pulse can be continuously measured accurately.

  Moreover, the power consumption of the measuring device 1 can be reduced by reducing the calculation amount. As a result, for example, the measurement device 1 can be mounted on a subject for a long time without performing charging or battery replacement, and measurement can be performed.

  In the above description, the case where the downsampling ratio is set to 1/16 has been described. However, if the sampling frequency of the measurement signal after downsampling is 8.0 Hz or more, the pulse wave frequency for the pulse rate of 240 bpm is set. It is possible to measure. Therefore, when the sampling frequency of the measurement signal is 200 Hz, the downsampling ratio can be reduced to 1/25.

  For example, FIG. 11 shows an example in which the downsampling ratio is set to 1/24. In this case, it is possible to detect up to a frequency component of 4.17 Hz (= 200 Hz × 1/24/2) using the measurement signal after downsampling, as indicated by a range R1 ′ in FIG.

<3. Second Embodiment>
Next, a second embodiment of the present technology will be described with reference to FIGS. The second embodiment differs from the first embodiment in a body motion detection method and a pulse wave frequency measurement method. In the second embodiment, measurement light having two types of wavelengths, measurement light 1 and measurement light 2, is used.

  As the measurement light 1, for example, blue measurement light having a wavelength of 470 nm or green measurement light having a wavelength of 530 nm is used. Hereinafter, a case where blue measurement light is used will be described as an example. The measurement signal 1 is measured using the measurement light 1.

  For the measurement light 2, for example, yellow measurement light having a wavelength of 585 nm is used. The measurement signal 2 is measured using the measurement light 2.

  Further, the sampling frequency of the measurement signal 1 and the measurement signal 2 is, for example, 200 to 220 Hz. Hereinafter, the case where the sampling frequency of the measurement signal 1 and the measurement signal 2 is 200 Hz will be described as an example.

{Configuration example of arithmetic processing unit 26b}
In the second embodiment, an arithmetic processing unit 26b of FIG. 12 is used in the measurement apparatus 1 instead of the arithmetic processing unit 26a of FIG. The arithmetic processing unit 26b includes decimation filters 301a and 301b, BPFs (bandpass filters) 302a and 302b, an autocovariance function estimation unit 303, a linear prediction filter 304, a BPF (bandpass filter) 305, an arithmetic unit 306, and a combined vector calculation. A unit 307, a determination unit 308, a selection unit 309, a DFT (discrete Fourier transform) unit 310, a band limiting unit 311, a peak detection unit 312, a calculation unit 313, and a storage unit 314.

  The body motion signal extraction unit 331 is configured by the self-covariance function estimation unit 303 and the linear prediction filter 304. The BPF 302a, the BPF 305, the calculation unit 306, and the body motion signal extraction unit 331 constitute a pulse wave signal extraction unit 332. The combined vector calculation unit 307 and the determination unit 308 constitute a body motion detection unit 333. The DFT unit 310, the band limiting unit 311, and the peak detection unit 312 constitute a frequency detection unit 334. The selection unit 309, the calculation unit 313, the storage unit 314, and the frequency detection unit 334 constitute a measurement unit 335.

  The decimation filters 301a and 301b have the same function as the decimation filter 101 in FIG. The BPFs 302a and 302b have the same function as the BPF 102 in FIG. The autocovariance function estimation unit 303 has the same function as the autocovariance function estimation unit 103 of FIG. The linear prediction filter 304 has the same function as the linear prediction filter 104 in FIG. The BPF 305 has the same function as the BPF 105 in FIG. The calculation unit 306 has the same function as the calculation unit 106 of FIG. The DFT unit 310 has the same function as the DFT units 107a and 107b in FIG. The band limiter 311 and the peak detector 312 realize the same function as the peak detectors 108a and 108b in FIG. The calculation unit 313 has the same function as the calculation unit 112 in FIG.

  The decimation filter 301a performs downsampling of the measurement signal 1. The decimation filter 301a supplies the down-sampled measurement signal 1 to the BPF 302a and the BPF 305.

  The decimation filter 301b performs downsampling of the measurement signal 2. The decimation filter 301b supplies the measurement signal 2 after downsampling to the BPF 302b.

  The BPF 302a extracts the frequency band N component of the measurement signal 1, and uses the extracted measurement signal including the frequency band N component (hereinafter referred to as the band signal 1N) as the self-covariance function estimation unit 303 and the linear prediction filter 304. , And the combined vector calculation unit 307.

  The BPF 302b extracts a component of a predetermined frequency band N of the measurement signal 2, and supplies a measurement signal including the extracted component of the frequency band N (hereinafter referred to as a band signal 2N) to the combined vector calculation unit 307.

  Similar to the self-covariance function estimation unit 103 in FIG. 6, the auto-covariance function estimation unit 303 estimates the self-covariance function of the body motion signal included in the band signal 1N and supplies the estimation result to the linear prediction filter 304. To do.

  Similar to the linear prediction filter 104 in FIG. 6, the linear prediction filter 304 generates an AR model of a body motion signal using the estimated autocovariance function, and generates a band signal based on the prediction result of the body motion signal. A body motion signal is extracted from 1N. The linear prediction filter 304 supplies the extracted body motion signal to the calculation unit 306.

  The BPF 305 extracts a component of the frequency band W of the measurement signal 1 and supplies a signal including the extracted component of the frequency band W (hereinafter referred to as a band signal 1W) to the calculation unit 306 and the selection unit 309.

  The calculation unit 306 adds the band signal W1 and the inverse signal of the body motion signal to obtain a difference between the band signal W1 and the body motion signal. The calculation unit 306 supplies a pulse wave signal, which is a difference signal between the band signal W1 and the body motion signal, to the selection unit 309.

  The combined vector calculation unit 307 calculates a combined vector of the band signal 1N and the band signal 2N. The combined vector calculation unit 307 supplies the calculation result of the combined vector to the determination unit 308.

  The determination unit 308 determines the presence / absence of obstacle body movement based on the calculation result of the combined vector. The determination unit 308 supplies the determination result of the presence / absence of obstacle body movement to the selection unit 309.

  The selection unit 309 selects a signal used for pulse measurement (hereinafter referred to as a pulse measurement signal) from the pulse wave signal and the band signal 1W based on the determination result of the presence or absence of obstacle body movement. The selection unit 309 supplies the selected pulse measurement signal to the DFT unit 310.

  Similar to the DFT units 107 a and 107 b in FIG. 6, the DFT unit 310 performs DFT of the pulse measurement signal and supplies the analysis result of the frequency of the pulse measurement signal to the band limiting unit 311.

  The band limiting unit 311 limits the frequency band for detecting the peak frequency based on the previous detection value of the pulse wave frequency stored in the storage unit 314. The band limiting unit 311 supplies the peak detection unit 312 with the analysis result of the frequency of the pulse measurement signal and information indicating the frequency band for detecting the peak frequency.

  The peak detector 312 detects the peak frequency of the pulse measurement signal in the same manner as the peak detectors 108a and 108b in FIG. This peak frequency becomes a measured value of the pulse wave frequency. The peak detection unit 312 supplies the measurement value of the pulse wave frequency to the calculation unit 313 and stores it in the storage unit 314.

  The calculation unit 313 calculates the pulse rate based on the pulse wave frequency. The calculation unit 313 outputs the calculated pulse rate to the outside as a measurement result.

  The memory | storage part 314 memorize | stores the measured value of the past pulse wave frequency.

{Second Embodiment of Pulse Measurement Processing}
Next, a second embodiment of the pulse measurement process executed by the measurement device 1 will be described with reference to the flowchart of FIG.

  In step S101, measurement signal acquisition is started in the same manner as in step S1 of FIG.

  In step S102, the decimation filters 301a and 301b down-sample the measurement signal. Specifically, the decimation filter 301a down-samples the measurement signal 1 at a predetermined ratio, and supplies the measurement signal 1 after down-sampling to the BPF 302a and the BPF 305. The decimation filter 301b downsamples the measurement signal 2 at a predetermined ratio, and supplies the measurement signal 2 after downsampling to the BPF 302b.

  In step S103, the BPF 302a, 302b and the BPF 305 limit the frequency band of the measurement signal. Specifically, the BPF 302a extracts the frequency band N component of the measurement signal 1 after downsampling. The BPF 302a supplies the extracted band signal 1N including the component of the frequency band N to the self-covariance function estimation unit 303, the linear prediction filter 304, and the combined vector calculation unit 307.

  The BPF 302b extracts the component of the frequency band N of the measurement signal 2 after downsampling. The BPF 302b supplies the band signal 2N including the extracted frequency band N component to the synthesized vector calculation unit 307.

  The BPF 305 extracts a component of the frequency band W of the measurement signal 1 after downsampling. The BPF 305 supplies the band signal 1W including the extracted component of the frequency band W to the calculation unit 306 and the selection unit 309.

  In step S104, a body motion signal is extracted from the band signal 1N by the same processing as in step S4 of FIG. The extracted body motion signal is supplied to the calculation unit 306.

  In step S105, the difference between the band signal 1W and the body motion signal is obtained by the same processing as in step S5 in FIG. 8, and a pulse wave signal is extracted. The extracted pulse wave signal is supplied to the selection unit 309.

  In step S106, the body motion detection unit 333 detects body motion. First, the combined vector calculation unit 307 calculates a combined vector of the band signal 1N and the band signal 2N. The combined vector is a vector having the sample value (amplitude value) of the band signal 1N and the sample value (amplitude value) of the band signal 2N at the same sampling time as components. The combined vector calculation unit 307 supplies the calculation result of the combined vector to the determination unit 308.

  The determination unit 308 determines the presence / absence of obstacle body movement based on the combined vector. Here, with reference to FIG. 14 thru | or FIG. 16, the determination method of the presence or absence of an obstruction body motion is demonstrated.

  14 and 15 are graphs showing examples of time-series changes in the signal level of the measurement signal. The horizontal axis of each graph indicates time, and the vertical axis indicates the value of the measurement signal.

  FIG. 14 shows the time series of the signal level of the measurement signal when the subject moves only the arm finger wearing the measuring device 1 in the period from about 40 seconds to about 80 seconds and is stationary in the other period. It shows a change. The uppermost graph shows an example in which blue measurement light having a wavelength of 470 nm is used. The second graph from the top shows an example when yellow measurement light having a wavelength of 585 nm is used. The third graph shows an example where green measurement light having a wavelength of 530 nm is used.

  Before the subject moves his / her finger, the measurement signals for the measurement lights of all wavelengths are not changed so much and are stable.

  Thereafter, when the subject moves his / her finger, the vibration width of the measurement signal with respect to the yellow measurement light becomes very large. However, although the values of the measurement signals for the measurement lights of other colors slightly increase, the vibration width does not change much.

  Even after the subject stops moving the finger, the vibration width of the measurement signal with respect to the yellow measurement light is reduced, but for a while the vibration width of the measurement signal with respect to the yellow measurement light is the vibration of the measurement signal with respect to the measurement light of other colors. The state larger than the width continues. In addition, for a while, the measurement signal values for the measurement lights of all wavelengths continue to be larger than when stationary.

  Thereafter, the measurement signals for the measurement lights of all wavelengths are returned to the stable state before the finger is moved.

  FIG. 15 shows a time-series change in the signal level of the measurement signal when the subject moves the entire arm on which the measurement apparatus 1 is worn during a period from about 78 seconds to about 118 seconds and is stationary during the other period. Is shown. The uppermost graph shows an example in which red measurement light having a wavelength of 660 nm is used. The second graph from the top shows an example when blue measurement light having a wavelength of 470 nm is used. The third graph from the top shows an example when yellow measurement light having a wavelength of 585 nm is used.

  As in the example of FIG. 14, before the subject moves his / her arm, the measurement signals for the measurement lights of all wavelengths are not changed so much and are stable.

  Thereafter, when the subject moves his arm, the value of the measurement signal for the measurement light of all wavelengths increases, and the vibration width increases.

  Even after the subject stops the movement of the arm, although the vibration width is small, the state of the high value of the measurement signal for the measurement light of all wavelengths continues.

  FIG. 16 shows an example of the distribution of the combined vector within one measurement period. The graph on the left shows an example of the distribution of the combined vector when the subject is stationary, and the graph on the right shows an example of the distribution of the combined vector when the subject is moving. The horizontal axis of the left and right graphs shows the sample value of the measurement signal (band signal 1N) for the blue measurement light, and the vertical axis shows the sample value of the measurement signal (band signal 2N) for the yellow measurement light.

  As described above with reference to FIGS. 14 and 15, the measurement signals for the measurement lights of all wavelengths are not changed so much and are stable before the occurrence of body movement. Therefore, as shown in the graph on the left side of FIG. 16, it is assumed that the distribution of the combined vector when there is no obstacle movement is within a normal zone 351 that is a predetermined range surrounded by a dotted line.

  On the other hand, when body motion occurs, the value of the measurement signal increases and the vibration width increases. However, the change in the measurement signal differs depending on the wavelength of the measurement light as well as the type and size of the body movement. That is, as shown in the example of FIG. 14, depending on the type of body movement, there is a case where only the measurement signal with respect to the measurement light of a specific wavelength reacts greatly.

  Accordingly, as shown in the graph on the right side of FIG. 16, the distribution of the combined vector when there is an obstacle movement is larger than the distribution of the combined vector at rest, and protrudes from the normal zone 351. In addition, since the change in the measurement signal may vary depending on the wavelength of the measurement light, the combined vector may deviate significantly from the upward straight line passing through the origin.

  Therefore, for example, the determination unit 308 determines that there is no obstacle movement when the distribution of the combined vector falls within the normal zone 351. On the other hand, the determination unit 308 determines that there is an obstacle movement when the distribution of the combined vector does not fall within the normal zone 351. The determination unit 308 supplies the determination result to the selection unit 309.

  In step S107, the selection unit 309 selects a signal (pulse measurement signal) used for pulse measurement based on the detection result of body movement. Specifically, when it is determined that there is an obstacle body motion, the selection unit 309 selects the pulse wave signal from the calculation unit 306 as a pulse measurement signal and supplies the pulse measurement signal to the DFT unit 310. On the other hand, if it is determined that there is no obstacle movement, the selection unit 309 selects the band signal 1W from the BPF 305 as a pulse measurement signal and supplies the pulse measurement signal to the DFT unit 310.

  In step S108, the frequency detector 334 detects the pulse wave frequency. Specifically, the DFT unit 310 performs frequency analysis of the pulse measurement signal by the same processing as step S6 in FIG. 8 and supplies the analysis result to the band limiting unit 311.

  The band limiting unit 311 limits the frequency band for detecting the peak frequency based on the detection value of the previous pulse wave frequency stored in the storage unit 314 by the same process as step S6 in FIG. The band limiting unit 311 supplies the peak detection unit 312 with the analysis result of the frequency of the pulse measurement signal and information indicating the frequency band for detecting the peak frequency.

  The peak detector 312 detects the peak frequency of the pulse measurement signal by the same process as step S6 of FIG. This peak frequency becomes a measured value of the pulse wave frequency.

  Therefore, when it is determined that there is an obstruction body motion, the peak frequency of the pulse wave signal becomes a measurement value of the current pulse wave frequency. On the other hand, when it is determined that there is no obstacle movement, the peak frequency of the band signal 1W is the measured value of the current pulse wave frequency.

  The peak detection unit 312 supplies information indicating the pulse wave frequency to the calculation unit 112 and causes the storage unit 113 to store the information.

  In step S109, the pulse rate is calculated in the same manner as in step S10 of FIG.

  In step S110, the measurement result is output in the same manner as in step S11 of FIG.

  Thereafter, the process returns to step S102, and the processes after step S102 are repeatedly executed.

  In this way, as in the first embodiment, the influence of body movement can be removed with a small amount of calculation, and the pulse wave and pulse of the subject can be accurately measured.

<4. Third Embodiment>
Next, a third embodiment of the present technology will be described with reference to FIGS. 17 to 19. The second embodiment differs from the first embodiment and the second embodiment in the body motion signal extraction method and the body motion detection method. In the third embodiment, for example, measurement light having one type of wavelength is used as in the first embodiment.

{Configuration example of arithmetic processing unit 26c}
In the third embodiment, an arithmetic processing unit 26c in FIG. 17 is used in the measurement apparatus 1 instead of the arithmetic processing unit 26a in FIG. 6 and the arithmetic processing unit 26b in FIG. The arithmetic processing unit 26c includes a decimation filter 501, a BPF (band pass filter) 502, a variable notch filter 503, an autocovariance function estimation unit 504, a linear prediction filter 505, a calculation unit 506, a body motion detection unit 507, a selection unit 508, A DFT (discrete Fourier transform) unit 509, a band limiting unit 510, a peak detection unit 511, a calculation unit 512, and a storage unit 513 are included.

  The variable notch filter 503, the autocovariance function estimation unit 504, and the linear prediction filter 505 constitute a body motion signal extraction unit 531. The BPF 502, the calculation unit 506, and the body motion signal extraction unit 531 constitute a pulse wave signal extraction unit 532. The DFT unit 509, the band limiting unit 510, and the peak detection unit 511 constitute a frequency detection unit 533. The selection unit 508, the calculation unit 512, the storage unit 513, and the frequency detection unit 533 constitute a measurement unit 534.

  The decimation filter 501 has the same function as the decimation filter 101 in FIG. The BPF 502 has the same function as the BPF 105 in FIG. The autocovariance function estimation unit 504 has the same function as the autocovariance function estimation unit 103 of FIG. The linear prediction filter 505 has the same function as the linear prediction filter 104 of FIG. The calculation unit 506 has the same function as the calculation unit 106 of FIG. The selection unit 508 has the same function as the selection unit 309 in FIG. The DFT unit 509 has the same function as the DFT units 107a and 107b in FIG. The band limiting unit 510 has the same function as the band limiting unit 311 in FIG. The peak detection unit 511 has the same function as the peak detection unit 312 in FIG. The calculation unit 512 has the same function as the calculation unit 112 in FIG.

  The decimation filter 501 performs downsampling of the measurement signal. The decimation filter 501 supplies the measurement signal after downsampling to the BPF 502.

  The BPF 502 extracts the component of the frequency band W of the measurement signal and supplies the band signal W including the extracted component of the frequency band W to the variable notch filter 503, the calculation unit 506, and the selection unit 508.

  The variable notch filter 503 is a zero phase filter with a variable attenuation band. The variable notch filter 503 sets the attenuation band so as to include the previous measurement value of the pulse wave frequency stored in the storage unit 513, and attenuates the component of the attenuation band among the frequency components of the band signal W. The variable notch filter 503 supplies the attenuated signal (hereinafter referred to as a band signal W ′) to the self-covariance function estimation unit 504 and the linear prediction filter 505.

  Since the variable notch filter 503 is a zero phase filter, the time axis information such as the body motion component is held as it is in the attenuated band signal W ′.

  Similar to the self-covariance function estimation unit 103 in FIG. 6, the auto-covariance function estimation unit 504 estimates the self-covariance function of the body motion signal included in the band signal W ′ and sends the estimation result to the linear prediction filter 505. Supply.

  Similar to the linear prediction filter 104 in FIG. 6, the linear prediction filter 505 generates an AR model of the body motion signal using the estimated autocovariance function, and generates a band signal based on the prediction result of the body motion signal. A body motion signal is extracted from W ′. The linear prediction filter 505 supplies the extracted body motion signal to the calculation unit 506.

  The calculation unit 506 adds the band signal W and the inverse signal of the body motion signal to obtain a difference between the band signal W and the body motion signal. The calculation unit 506 supplies a pulse wave signal, which is a difference signal between the band signal W and the body motion signal, to the selection unit 508.

  The body motion detection unit 507 detects body motion by a predetermined method and determines the presence or absence of obstacle body motion. The body motion detection unit 507 supplies the determination result of the presence or absence of the obstacle body motion to the selection unit 508.

  The selection unit 508 selects a pulse measurement signal from the pulse wave signal and the band signal W based on the determination result of the presence or absence of obstacle body movement. The selection unit 508 supplies the selected pulse measurement signal to the DFT unit 509.

  Similar to the DFT units 107 a and 107 b in FIG. 6, the DFT unit 509 performs DFT of the pulse measurement signal and supplies the analysis result of the frequency of the pulse measurement signal to the band limiting unit 510.

  Band limiting section 510 limits the frequency band for detecting the peak frequency based on the previous detected pulse wave frequency value stored in storage section 513. The band limiting unit 510 supplies the peak detection unit 511 with the frequency analysis result of the pulse measurement signal and information indicating the frequency band for detecting the peak frequency.

  The peak detector 511 detects the peak frequency of the pulse measurement signal in the same manner as the peak detectors 108a and 108b in FIG. This peak frequency becomes a measured value of the pulse wave frequency. The peak detection unit 511 supplies the measurement value of the pulse wave frequency to the calculation unit 512 and stores it in the storage unit 513.

  Calculation unit 512 calculates the pulse rate based on the pulse wave frequency. The calculation unit 512 outputs the calculated pulse rate to the outside as a measurement result.

  The memory | storage part 513 memorize | stores the measured value of the past pulse wave frequency.

{Third embodiment of pulse measurement processing}
Next, a third embodiment of the pulse measurement process executed by the measurement device 1 will be described with reference to the flowchart of FIG.

  In step S301, measurement signal acquisition is started in the same manner as in step S1 of FIG.

  The top graph in FIG. 19 shows an example of the waveform of the measurement signal. The horizontal axis of the graph indicates time, and the vertical axis indicates the amplitude value.

  In step S302, the measurement signal is down-sampled in the same manner as in step S2 of FIG. Then, the measurement signal after downsampling is supplied to the BPF 502.

  In step S303, the BPF 502 limits the frequency band of the measurement signal. Specifically, the BPF 502 extracts the frequency band W component of the measurement signal after downsampling. The BPF 502 supplies the band signal W including the extracted component of the frequency band W to the variable notch filter 503, the calculation unit 506, and the selection unit 508.

  The second graph from the top in FIG. 19 shows an example of the frequency distribution of the band signal W after limiting the frequency band of the measurement signal shown in the top graph. The horizontal axis of the graph indicates the frequency, and the vertical axis indicates the spectral intensity. In this example, a pulse wave spectrum appears near 1.6 Hz, and a body motion component spectrum appears at 1.5 Hz or lower. Further, there is almost no frequency component of 1.8 Hz or higher.

  In step S304, the variable notch filter 503 attenuates the frequency component in the vicinity of the previous measurement value of the pulse wave frequency among the frequency components of the measurement signal (band signal W) after band limitation. Specifically, the variable notch filter 503 reads the measurement value of the previous pulse wave frequency from the storage unit 513. The variable notch filter 503 sets, for example, a predetermined band including a measurement value of the previous pulse wave frequency as an attenuation band. For example, a predetermined band centered on the previous measurement value of the pulse wave frequency is set as the attenuation band.

  The variable notch filter 503 attenuates the component of the set attenuation band among the frequency components of the band signal W. The variable notch filter 503 supplies the attenuated band signal W ′ to the self-covariance function estimation unit 504 and the linear prediction filter 505.

  Since the pulse wave frequency does not change abruptly, it is assumed that most of the pulse wave component is included in the frequency component attenuated by the variable notch filter 503. Accordingly, the pulse signal component is hardly included in the band signal W ′.

  In step S305, a body motion signal is extracted from the band signal W ′ by the same processing as in step S4 of FIG. The extracted body motion signal is supplied to the calculation unit 506.

  As described above, the band signal W ′ contains almost no pulse wave component. Therefore, unlike the embodiment described above, the pulse signal component is hardly included in the body motion signal regardless of the amount of the body motion component included in the measurement signal.

  The third graph from the top in FIG. 19 shows an example of the frequency distribution of the body motion signal extracted from the band signal W shown in the second graph. The horizontal axis of the graph indicates the frequency, and the vertical axis indicates the spectral intensity. As shown in this graph, the spectrum of the pulse wave included in the band signal W disappears in the body motion signal.

  In step S306, the difference between the band signal W and the body motion signal is obtained by the same processing as in step S5 of FIG. 8, and a pulse wave signal is extracted. The extracted pulse wave signal is supplied to the selection unit 508.

  As described above, since the pulse wave component is hardly included in the body motion signal, the pulse wave component is surely included in the pulse wave signal.

  The fourth graph from the top in FIG. 19 shows an example of the frequency component of the pulse wave signal obtained by taking the difference between the band signal W shown in the second graph from the top and the body motion signal in the third graph. ing. The horizontal axis of the graph indicates the frequency, and the vertical axis indicates the spectral intensity. As shown in this graph, a pulse wave component is extracted from the band signal W, and other components are attenuated in number. Therefore, the pulse wave component is very large compared to the body motion component.

  In step S307, the body motion detection unit 507 detects body motion. For example, the body motion detection unit 507 detects body motion based on a detection signal from a sensor (not shown) such as a triaxial acceleration sensor or a gyro sensor included in the measurement device 1. Then, the body motion detection unit 507 determines the presence or absence of the obstacle body motion based on the detection result of the body motion. The body motion detection unit 507 supplies the determination result of the presence or absence of the obstacle body motion to the selection unit 508.

  In step S308, as in the process of step S107 in FIG. 13, a signal (pulse measurement signal) used for pulse measurement is selected based on the detection result of body movement.

  In step S309, the pulse wave frequency is detected in the same manner as in step S108 of FIG.

  The fifth graph from the top in FIG. 19 shows the frequency of the pulse wave signal after the detection range is limited in step S309 when the pulse wave signal in the fourth graph from the top is selected as the pulse measurement signal. Distribution is shown. The horizontal axis of the graph indicates the frequency, and the vertical axis indicates the spectral distribution.

  The bottom graph of FIG. 19 shows the waveform of the pulse wave signal after subtracting the body motion signal having the frequency distribution of the third graph from the measurement signal of the first graph from the top. The horizontal axis of the graph indicates time, and the vertical axis indicates the amplitude value.

  In step S310, the pulse rate is calculated in the same manner as in step S109 of FIG.

  In step S311, the measurement result is output in the same manner as in step S110 of FIG.

  Thereafter, the process returns to step S302, and the processes after step S302 are repeatedly executed.

  In this way, as in the first and second embodiments, the influence of body movement can be removed with a small amount of computation, and the pulse wave and pulse of the subject can be accurately measured. .

  Further, in the third embodiment, one BPF can be reduced as compared with the first embodiment and the second embodiment. Furthermore, in the third embodiment, a pulse wave signal obtained by extracting a pulse wave component with high accuracy can be generated.

<5. Modification>
Hereinafter, modifications of the above-described embodiment of the present technology will be described.

{Variant related to body motion detection method}
The body motion detection method described above is an example, and other methods may be employed.

  For example, as described above with reference to FIGS. 14 and 15, when body movement occurs, the signal level of the measurement signal changes greatly. Therefore, for example, the body motion detection unit may detect body motion based on a change in the signal level (for example, amplitude value, amplitude width, etc.) of the measurement signal. For example, the body motion detection unit 133 or the body motion detection unit 333 may determine that there is an obstacle body motion when the amount of change in the signal level of the measurement signal is equal to or greater than a predetermined threshold.

  Further, as described above with reference to FIGS. 14 and 15, the tendency of the measurement signal to vary greatly differs depending on the wavelength of the measurement light and the type of body movement. Therefore, for example, when measuring light having two or more types of wavelengths is used, the body motion detection unit determines that obstacle body motion has occurred if at least one of a plurality of measurement signals for each measurement light satisfies a predetermined condition. You may make it do.

  Further, FIG. 20 schematically shows the measurement signal before limiting the frequency band and the envelope of the measurement signal after limiting. As shown in this figure, when the band is limited, the rise and fall of the measurement signal change become dull. Therefore, there may be a delay in detection of body movement or erroneous detection of body movement. Therefore, for example, the body motion detection unit may detect the body motion based on the measurement signal before removing the DC component and before limiting the frequency band or performing downsampling.

  Further, body motion may be detected by combining a plurality of body motion detection methods.

{Variant relating to pulse wave frequency detection method}
In the above description, an example in which one of the peak frequency of the pulse wave signal and the band signal W (or the band signal 1W) is selected as the pulse wave frequency based on the detection result of the body motion is shown. The pulse wave frequency may be measured by this method.

  For example, in the frequency distribution of the measurement signal, if there is a peak whose spectrum intensity is greater than or equal to a predetermined threshold in a predetermined band centered on the previous measurement value of the pulse wave frequency, the frequency corresponding to that peak is pulsed. You may make it measure as a wave frequency.

  In the third embodiment, the pulse wave component is surely included in the pulse wave signal. Therefore, for example, it is possible to omit the body motion detection process and always measure the peak frequency of the pulse wave signal as the pulse wave frequency.

{Variant related to AR model of body motion signal}
For example, the order of the AR model of the linear prediction filter may be changed based on the level (for example, amplitude value) of the measurement signal (band signal W or band signal 1W).

  Specifically, when the ratio of the body motion component to the pulse wave component is high, the accuracy of separating the body motion component and the pulse wave component of the linear prediction filter does not decrease much even if the order of the AR model of the linear prediction filter is increased. Also, by increasing the order of the AR model of the linear prediction filter, the noise separation accuracy of the linear prediction filter is improved. On the other hand, when the ratio of the body motion component to the pulse wave component is low, the accuracy of separation of the body motion component and the pulse wave component of the linear prediction filter decreases when the order of the AR model of the linear prediction filter is increased.

  Further, it is assumed that the higher the level of the measurement signal, the larger the body motion component included in the measurement signal and the higher the ratio of the body motion component to the pulse wave component.

  Therefore, for example, the linear prediction filter may increase the order of the AR model as the level of the measurement signal increases, and decrease the order of the AR model as the level of the measurement signal decreases.

  Further, the AR model of the body motion signal may be generated by a method other than the Yule-Walker method.

{Modifications related to downsampling}
For example, by extracting a component of a predetermined band (for example, 1 to 2 octaves) centered on the previous pulse wave frequency by BPF and shifting the frequency of the extracted signal, the ratio of down-sampling the measurement signal is further increased. It is possible to lower. A specific example will be described with reference to FIG. 21 is the same as FIG.

  For example, first, the component of the frequency band of the detection range R2 is extracted from the measurement signal by the BPF. Next, the frequency band of the extracted signal is shifted to the frequency band R2 ″ in the lower diagram. Thereby, the detectable frequency component of the measurement signal after downsampling is not less than the range R1 ″ that is narrower than the range R1 If it can be ensured, the pulse wave frequency can be detected. Therefore, the downsampling ratio of the measurement signal can be further reduced.

  Further, for example, the sampling frequency of the measurement signal may be lowered within a range in which the pulse wave frequency can be measured without performing downsampling.

{Modifications related to measurement results}
In the above description, an example in which the pulse rate is output as a measurement result has been described. However, for example, a pulse wave frequency can be output as a measurement result.

  For example, it is also possible to output a pulse wave signal as a measurement result. Alternatively, one of the pulse wave signal and the band signal W (or band signal 1W) is selected based on the result of body motion detection, and the selected signal is output as a pulse wave measurement result. Also good. In any case, it is desirable to output the signal after removing the noise.

  Furthermore, two or more types of pulse rate, pulse wave frequency, and pulse wave signal (or band signal W or band signal 1W) may be output as measurement results.

  Further, for example, a body motion signal may be included in the measurement result.

{Other variations}
The various numerical values (for example, various frequencies and wavelengths, the number of samples, the ratio of downsampling, etc.) shown in the above description are examples thereof, and numerical values other than those listed above can be adopted.

  Further, instead of the measurement unit 135 of FIG. 6, the measurement unit 335 of FIG. 12 or the measurement unit 534 of FIG. 17 may be used. Conversely, instead of the measurement unit 335 in FIG. 12 or the measurement unit 534 in FIG. 17, the measurement unit 135 in FIG. 6 may be used.

  Further, as described above with reference to FIGS. 14 and 15, the change in the waveform of the measurement signal varies depending on the wavelength of the measurement light and the type of body movement. For example, the type of body movement may be classified based on the difference in waveform change.

  Furthermore, the processing of each step in the flowcharts of FIGS. 8, 13, and 18 can be appropriately changed in order or executed in parallel.

  It is also possible to configure the measuring device 1 by a system composed of a plurality of devices. For example, a part or all of the arithmetic processing unit 26 may be provided in a device different from the device worn on the subject, and measurement signals may be transmitted between the devices by wireless communication or wired communication.

  Furthermore, in the above description, an example in which the measurement device is mounted on a person's arm has been shown, but the present technology can also be applied to a measurement device mounted on a portion other than the arm. Also, the shape of the measuring device can be a shape other than the wristband type described above.

  Moreover, this technique is applicable also when measuring the pulse wave and pulse of living organisms other than a person.

{Example of computer configuration}
The series of processes described above can be executed by hardware or can be executed by software. When a series of processing is executed by software, a program constituting the software is installed in the computer. Here, the computer includes, for example, a general-purpose personal computer capable of executing various functions by installing various programs by installing a computer incorporated in dedicated hardware.

  FIG. 22 is a block diagram illustrating an example of a hardware configuration of a computer that executes the above-described series of processes using a program.

  In a computer, a CPU (Central Processing Unit) 701, a ROM (Read Only Memory) 702, and a RAM (Random Access Memory) 703 are connected to each other by a bus 704.

  An input / output interface 705 is further connected to the bus 704. An input unit 706, an output unit 707, a storage unit 708, a communication unit 709, and a drive 710 are connected to the input / output interface 705.

  The input unit 706 includes a keyboard, a mouse, a microphone, and the like. The output unit 707 includes a display, a speaker, and the like. The storage unit 708 includes a hard disk, a nonvolatile memory, and the like. The communication unit 709 includes a network interface. The drive 710 drives a removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory.

  In the computer configured as described above, the CPU 701 loads the program stored in the storage unit 708 to the RAM 703 via the input / output interface 705 and the bus 704 and executes the program, for example. Is performed.

  The program executed by the computer (CPU 701) can be provided by being recorded on a removable medium 711 as a package medium, for example. The program can be provided via a wired or wireless transmission medium such as a local area network, the Internet, or digital satellite broadcasting.

  In the computer, the program can be installed in the storage unit 708 via the input / output interface 705 by attaching the removable medium 711 to the drive 710. Further, the program can be received by the communication unit 709 via a wired or wireless transmission medium and installed in the storage unit 708. In addition, the program can be installed in advance in the ROM 702 or the storage unit 708.

  The program executed by the computer may be a program that is processed in time series in the order described in this specification, or in parallel or at a necessary timing such as when a call is made. It may be a program for processing.

  In this specification, the system means a set of a plurality of components (devices, modules (parts), etc.), and it does not matter whether all the components are in the same housing. Accordingly, a plurality of devices housed in separate housings and connected via a network and a single device housing a plurality of modules in one housing are all systems. .

  Furthermore, the embodiments of the present technology are not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present technology.

  For example, the present technology can take a configuration of cloud computing in which one function is shared by a plurality of devices via a network and is jointly processed.

  In addition, each step described in the above flowchart can be executed by being shared by a plurality of apparatuses in addition to being executed by one apparatus.

  Further, when a plurality of processes are included in one step, the plurality of processes included in the one step can be executed by being shared by a plurality of apparatuses in addition to being executed by one apparatus.

  Moreover, the effect described in this specification is an illustration to the last, and is not limited, There may exist another effect.

  Furthermore, the embodiments of the present technology are not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present technology.

  For example, this technique can also take the following structures.

(1)
Body motion including a component caused by body motion from a first band signal including a component of a first frequency band of a first measurement signal acquired by irradiating a portion including a pulse with light of a first wavelength A body motion signal extraction unit for extracting a signal;
A measuring device comprising: an arithmetic unit that generates a pulse wave signal that is a difference signal between a second band signal including a second frequency band component of the first measurement signal and the body motion signal.
(2)
The measurement apparatus according to (1), wherein the body motion signal extraction unit predicts and extracts the body motion signal using an autoregressive model.
(3)
The measurement device according to (2), wherein the body motion signal extraction unit generates the autoregressive model within a range from the fifth order to the twelfth order using the Yule-Walker method.
(4)
The measurement device according to (3), wherein the body motion signal extraction unit sets the order of the autoregressive model based on a level of the first band signal.
(5)
A body motion detector for detecting the body motion;
(1) to (4), further comprising: a measurement unit that measures a pulse wave frequency based on a signal selected from the pulse wave signal and the second band signal based on the detection result of the body motion. The measuring apparatus in any one of.
(6)
The body motion signal extraction unit extracts the body motion signal from a signal obtained by attenuating a frequency component of a band including a measurement value of a previous pulse wave frequency among frequency components of the first band signal. The measuring device according to 5).
(7)
The measuring unit is
A frequency detector that detects a first peak frequency that is a peak frequency of the pulse wave signal and a second peak frequency that is a peak frequency of the second band signal;
A selection unit that selects the pulse wave frequency from the first peak frequency and the second peak frequency based on at least one of the detection result of the body motion and the previous measurement value of the pulse wave frequency; The measuring apparatus according to (5), provided.
(8)
The measurement device according to (7), wherein the frequency detection unit limits a frequency band in which the first peak frequency and the second peak frequency are detected based on a measurement value of a previous pulse wave frequency.
(9)
The frequency detection unit detects the first peak frequency based on a result of Fourier transform of the pulse wave signal obtained by padding a sample having a value of 0, and the second band in which the sample having a value of 0 is padded. The measurement device according to (7) or (8), wherein the second peak frequency is detected based on a result of Fourier transform of the signal.
(10)
The measuring unit is
A selection unit that selects the pulse wave signal or the second band signal based on the detection result of the body movement;
The measurement device according to (5), further comprising: a frequency detection unit that detects a peak frequency of the signal selected by the selection unit as the pulse wave frequency.
(11)
The measurement device according to (10), wherein the frequency detection unit limits a frequency band in which the peak frequency is detected based on a measurement value of a previous pulse wave frequency.
(12)
The frequency detection unit detects the peak frequency based on a result obtained by performing Fourier transform on a signal obtained by padding a signal with a value of 0 to the signal selected by the selection unit. (10) or (11) Measuring device.
(13)
The measurement device according to any one of (5) to (12), wherein the body motion detection unit detects the body motion based on a frequency distribution of the body motion signal.
(14)
The body motion detection unit includes a third band signal including a component of the first frequency band of a second measurement signal acquired by irradiating a portion including the pulse with light of a second wavelength, and the The measurement device according to any one of (5) to (13), wherein the body motion is detected based on a distribution of a combined vector with the first band signal.
(15)
The body motion detection unit is based on the change in the first measurement signal and the change in the second measurement signal acquired by irradiating the portion including the pulse with the light having the second wavelength. The measurement device according to any one of (5) to (14), wherein body movement is detected.
(16)
The measurement device according to any one of (5) to (15), wherein the measurement unit calculates a pulse rate based on the pulse wave frequency.
(17)
A first filter for extracting the first band signal from the first measurement signal;
A second filter for extracting the second band signal from the first measurement signal,
The second frequency band includes a pulse wave frequency range to be measured, and the maximum value of the second frequency band is larger than the maximum value of the first frequency band. (1) to (16) The measuring apparatus in any one of.
(18)
A filter for extracting the first band signal from the first measurement signal;
The first frequency band is the same as the second frequency band, and includes a range of pulse wave frequencies to be measured,
The measurement device according to any one of (1) to (16), wherein the first band signal and the second band signal are the same signal.
(19)
A body motion signal including a component caused by body motion from a first band signal including a component of a first frequency band of a first measurement signal acquired by irradiating a portion including a pulse with light of a predetermined wavelength A body motion signal extraction step for extracting
A calculation method including: a calculation step of generating a pulse wave signal that is a difference signal between a second band signal including a component of a second frequency band of the first measurement signal and the body motion signal.

  DESCRIPTION OF SYMBOLS 1 Measuring apparatus, 11 Main-body part, 22, 22a thru | or 22c LED, 23 Light-receiving IC, 26, 26a thru | or 26c Arithmetic processing part, 51 LED driver, 53 Light receiving element, 54 AD converter, 101 Decimation filter, 102 BPF, 103 Self Variance function estimation unit, 104 linear prediction filter, 105 BPF, 106 operation unit, 107a, 107b DFT unit, 108a, 108b peak detection unit, 109 DFT unit, 110 determination unit, 111 selection unit, 112 calculation unit, 131 body motion signal Extraction unit, 132 pulse wave signal extraction unit, 133 body motion detection unit, 134 frequency detection unit, 135 measurement unit, 301a, 301b decimation filter, 302a, 302b BPF, 303 self-covariance function estimation unit, 3 04 linear prediction filter, 305 BPF, 306 operation unit, 307 synthesized vector generation unit, 308 determination unit, 309 selection unit, 310 DFT unit, 311 band limiting unit, 312 peak detection unit, 313 calculation unit, 331 body motion signal extraction unit , 332 pulse wave signal extraction unit, 333 body motion detection unit, 334 frequency detection unit, 335 measurement unit, 501 decimation filter, 502 BPF, 503 variable notch filter, 504 self-covariance function estimation unit, 505 linear prediction filter, 506 calculation Unit, 507 body motion detection unit, 508 determination unit, 509 DFT unit, 510 band limiting unit, 511 peak detection unit, 512 calculation unit, 531 body motion signal extraction unit, 532 pulse wave signal extraction unit, 533 frequency detection unit, 534 Measuring unit

Claims (19)

Body motion including a component caused by body motion from a first band signal including a component of a first frequency band of a first measurement signal acquired by irradiating a portion including a pulse with light of a first wavelength A body motion signal extraction unit for extracting a signal;
A measuring device comprising: an arithmetic unit that generates a pulse wave signal that is a difference signal between a second band signal including a second frequency band component of the first measurement signal and the body motion signal. The measurement device according to claim 1, wherein the body motion signal extraction unit predicts and extracts the body motion signal using an autoregressive model. The measurement apparatus according to claim 2, wherein the body motion signal extraction unit generates the autoregressive model in a range from the fifth order to the twelfth order using a Yule-Walker method. The measuring apparatus according to claim 3, wherein the body motion signal extraction unit sets the order of the autoregressive model based on a level of the first band signal. A body motion detector for detecting the body motion;
The measurement part which measures a pulse wave frequency further based on the signal selected based on the detection result of the said body motion out of the said pulse wave signal and the said 2nd zone | band signal . The measuring device described in 1. The body motion signal extraction unit extracts the body motion signal from a signal obtained by attenuating a frequency component of a band including a measurement value of a previous pulse wave frequency among frequency components of the first band signal. 5. The measuring device according to 5. The measuring unit is
A frequency detector that detects a first peak frequency that is a peak frequency of the pulse wave signal and a second peak frequency that is a peak frequency of the second band signal;
A selection unit that selects the pulse wave frequency from the first peak frequency and the second peak frequency based on at least one of the detection result of the body motion and the previous measurement value of the pulse wave frequency; The measuring device according to claim 5 provided. The measurement device according to claim 7, wherein the frequency detection unit limits a frequency band in which the first peak frequency and the second peak frequency are detected based on a previous measurement value of the pulse wave frequency. The frequency detection unit detects the first peak frequency based on a result of Fourier transform of the pulse wave signal obtained by padding a sample having a value of 0, and the second band in which the sample having a value of 0 is padded. signal based on the result of Fourier transform of the measurement device according to claim 7 or 8 for detecting the second peak frequency. The measuring unit is
A selection unit that selects the pulse wave signal or the second band signal based on the detection result of the body movement;
The measurement device according to claim 5, further comprising: a frequency detection unit that detects a peak frequency of the signal selected by the selection unit as the pulse wave frequency. The measurement device according to claim 10, wherein the frequency detection unit limits a frequency band in which the peak frequency is detected based on a measurement value of a previous pulse wave frequency. Wherein the frequency detector, a signal value on the signal selected by the selection portion is padded samples of 0 on the basis of the result of Fourier transform, the measuring device according to claim 10 or 11 detects the peak frequency. The body motion detecting unit, based on the frequency distribution of the body movement signal, measuring apparatus according to any one of claims 5 to 12 for detecting the body motion. The body motion detection unit includes a third band signal including a component of the first frequency band of a second measurement signal acquired by irradiating a portion including the pulse with light of a second wavelength, and the based on the distribution of the synthesis vector of the first band signal, measuring apparatus according to any one of claims 5 to 13 for detecting the body motion. The body motion detection unit is based on the change in the first measurement signal and the change in the second measurement signal acquired by irradiating the portion including the pulse with the light having the second wavelength. measurement apparatus according to any one of claims 5 to 14 for detecting the body movement. The measuring part, based on the pulse wave frequency measuring apparatus according to any one of claims 5 to 15 to calculate the pulse rate. A first filter for extracting the first band signal from the first measurement signal;
A second filter for extracting the second band signal from the first measurement signal,
The second frequency band includes a range of pulse wave frequencies to be measured, and a maximum value of the second frequency band is larger than a maximum value of the first frequency band . The measuring device described in 1. A filter for extracting the first band signal from the first measurement signal;
The first frequency band is the same as the second frequency band, and includes a range of pulse wave frequencies to be measured,
The measuring device according to any one of the first of claims 1 to 16 band signal and the second band signals are the same signal. A body motion signal including a component caused by body motion from a first band signal including a component of a first frequency band of a first measurement signal acquired by irradiating a portion including a pulse with light of a predetermined wavelength A body motion signal extraction step for extracting
A calculation method including: a calculation step of generating a pulse wave signal that is a difference signal between a second band signal including a component of a second frequency band of the first measurement signal and the body motion signal.

Images